AskDocs

AskDocs MCP Server

python

6

GitHub Stars

python

Language

6 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "dymk-askdocs-mcp": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "--network=host",
        "--volume=${workspaceFolder}/docs:/docs",
        "ghcr.io/dymk/askdocs-mcp:latest"
      ],
      "env": {
        "ASKDOCS_OLLAMA_URL": "http://localhost:11434"
      }
    }
  }
}

You deploy and run an MCP server that powers semantic search over your technical PDFs using Ollama, enabling fast, context-aware answers drawn from your documents. This guide walks you through using the server, installing it locally, and configuring it to fit your workspace.

How to use

You run the MCP server with a client to connect your apps or tools. Start the server using the provided Docker command, then query it with natural language to search across your PDFs. The server exposes endpoints that let you list sources, ask questions against a document, and fetch page content when you need precise references.

How to install

Prerequisites you need before installation: a Docker runtime on your machine, and a workspace directory containing your PDFs and a configuration file.

Step 1: Prepare your project structure and configuration. Create a docs directory and a configuration file named askdocs-mcp.toml inside it with your document details.

Step 2: Run the MCP server with Docker. Use the following command to start the server and keep its data isolated to your project workspace.

# Start the MCP server using Docker
# Ensure you are in your project root so paths resolve correctly

docker run -it --rm --network=host -v ./docs:/docs ghcr.io/dymk/askdocs-mcp:latest

Configuration

Configure your project to point to the MCP server and define your documentation sources. The server expects a TOML configuration in the docs directory that lists each document, its description, and the path to the PDF file.

[[doc]]
name = "my_manual"
description = "My Product Manual"
path = "pdf/manual.pdf"

Optional global settings let you choose embedding and language models as needed.

## Security and environment

The server relies on an Ollama instance running at a configurable URL. By default, the Ollama URL is `http://localhost:11434`. You can override this with the environment variable `ASKDOCS_OLLAMA_URL`.

## Troubleshooting and notes

If you encounter connectivity issues, ensure Ollama is running and reachable at the configured URL. Verify Docker has access to the host network and that the `docs` directory contains a valid configuration and PDFs. Add the `.askdocs-cache` directory to your versioning ignore rules to keep the vector cache out of source control.

## Tools and capabilities

You can list sources, perform semantic searches, and fetch document pages using the provided tools.

## Available tools

### list\_docs

List all documentation sources available to the MCP server.

### ask\_docs

Search documentation using natural language queries and receive relevant results with citations.

### get\_doc\_page

Retrieve the full text from specific pages of a document for precise context.